Protein sequences deduced from gene sequences of diverse bacteria and archaebacteria are yielding a wealth of new knowledge on protein functions, interactions and evolution. Some novel findings were studied by concerted methods including directed mutagenesis, spectroscopic/enzymological analyses and computer analyses of sequence databases. Findings include:- A. Bacterial homologs of major "eukaryotic" protein families. Bacterial homologs of Serine/Threonine protein kinases were further investigated in cyanobacteria and archaebacteria. These are evidently relatively similar to the protein kinase C subfamily, but their phosphorylation specificity remains to be confirmed. B. Novel proton exchange mechanism in glutamate dehydrogenase. This enzyme differs from other NAD/NADP dependent dehydrogenases in lacking a catalytic histidine residue. From sequence conservation and site-directed mutagenesis of the E. coli enzyme, in comparison with the crystal structure of the Clostridium symbiosum enzyme, a triad of lysine residues was identified in the active site. These have novel protonation properties and evidently act in concert in dicarboxylate substrate binding, deprotonation of ammonium and other proton exchange steps of glutamate dehydrogenase catalysis. C. Sequence families in complex bacteria. New sequences from multicellular and differentiating bacteria, Streptomyces, Myxococcus and various cyanobacteria and archaebacteria, were investigated by global database sequence similarity searches. More than 50 percent of the classifiable protein sequences did not have counterparts in E. coli and other well-studied enteric bacteria, and many of these had eukaryotic homologs. Examples of low-complexity segments and multiple repeats are emerging with increasing frequency. Significance of project: Genome sequences from metabolically and morphogenetically diverse bacteria continue to provide a rich and cost-effective source of new discoveries on the molecular functions and evolution of proteins.